Design Of Automatic Monitoring System For Continuous Advanced Treatment Process Of Papermaking Wastewater And Development Of Soft-sensing Models For COD

The wastewater of pulp and papermaking industry has many problems like large emissions and serious pollutions, and the COD of effluent after sequencing batch activated sludge process is still above 150mg/L and do not reach the requirement of the new pulp and paper industrial water pollutant discharge standards(GB3544-2008). So it is necessary to add advanced treatment process for the secondary effluent. Photoelectrocatalytic oxidation technology has the characteristics of efficient, fast, and no selective degradation, which is widely researched and applied in a variety of wastewater treatment. According to the characteristics of the SBR effluent of papermaking wastewater, the photoelectrocatalyic oxidation technology based on our teams’ previous study was chosen for further processing wastewater. At present, one of reasons that the photoelectrocatalyic oxidation technology has not been used in actual industrial wastewater treatment is the lack of automatic control system for this technology. So the automatic monitoring system for continuous advanced treatment process of papermaking wastewater was designed based on the photoelectrocatalyic oxidation technology.A Siemens S7-200 PLC and Kingview6.53 software were used for designing the automatic monitoring system. In this system, the on-line monitoring of flow, liquid level and pH during the wastewater treatment process can be achieved and precisely controlled. What is more, the system had the functions of data recording and fault alarming, which provided reliable data for subsequent improvement of processing craft and optimization of monitoring system. Then the automatic monitoring system was tested and operated by simulation wastewater of Rhodamine B(RB). The results indicated that the whole system operated with safety and stability according to the designed program and control strategy.In order to realize on-line monitoring of the effluent COD, real-time parameters ORP, DO and pH during the photoelectrocatalyic oxidation reactions were used to build soft-sensing models for COD with the methods of Multiple Linear Regression(MLR), Multi-layer perceptron(MLP), Back propagation neural network(BPNN), Radial basis neural network(RBNN), Generalized regression neural network(GRNN), Extreme Learning Machine(ELM) and Support Vector Machine for Regression(SVR). After training and validation, the RBNN model was determined as the best soft-sensing model. Then another set of data was used to test the best model, the results showed that the model could well forecasted COD values during the papermaking wastewater advanced treatment process with photoelectrocatalyic oxidation technology.Finally, the COD soft-sensing model was applied in the wastewater of a raw wood pulping plant and a market pulp papermaking plant for testing the adaptability of the model. The result showed that the model could not be well applied to the wastewater of two paper mill and the model adaptability was not good. The result further explained the soft-sensing models’ great dependence on technology objects. The model should be rebuilt with new data, if it is used for the wastewater of other paper mill.